Research On Automatic Attitude Adjustment And Image Calibration Technology Of Surface Defect Detection System

Based on the analysis of the existing surface defect detection of precision optical components on the prototype, an improved scheme which can be used for defect detection system in large-scale automatic assemble line is raised. This program improves the instrument leveling mechanism and the image acquisition algorithms, as well as the solution on the part of the image intensity inhomogeneity problem, which makes the surface defect detection have the ability to auto-leveling of the sample surface and using the largest format industrial CCD camera. And it can be exclude non-uniform image background effects for weak signal extraction. It greatly accelerates the testing speed and efficiency of the defect detection system. This program not only greatly improves the overall performance, but also makes the defect detection system more suitably for the actual situation in which the surface defect detection process requires the various planar optical elements for industrial scale mass production.In the leveling system, this program analyzes the shortcomings of the original manual leveling scheme on precision optics surface defect detection system, proposes the use of an image acquisition system for machine vision and stepper motor control automatic leveling scheme on three fixtures. By redesigning the entire holding and leveling system, this program establishes automatic control programs and including the hardware-controlled software and the digital imaging processing software. It achieves the function of optical surface self-leveling feature. And defect detection greatly improves the speed of the preparatory stage and the overall detection efficiency. After the optimization and improvement program of the customization device, the system can be done in real-time leveling when it checks.In terms of image acquisition algorithms, this program analyzes the problem on the using of the original large-format CCD camera image acquisition. And for the reason causing the problem, adjustment algorithm is proposed on intensity differences of large-format dual-channel CCD image. This algorithm analyzes the differences in the characteristics of the image gray and establishes the gray matrix model. It also proposes an adaptively adjust the gradation method of dual-channel CCD. Experimental results show that the use of the algorithm can reduce the image mean gray differences to 0.70 from 3.64, solute the problems effectively on the use of large-format dual-channel CCD camera to capture images, let the system be significantly improved by the use of more large-format defective CCD camera detection speed detection system.In the image background non-uniformity, faces with the problem of encountering uneven gray and light ghosting due to the uneven illumination in actual use. After the common understanding of the domestic and non-uniform illumination image correction algorithm, this paper studies and analyses of the system of uneven image background, and then use the most appropriate background estimation method to solve such problems effectively. Experimental results show that the problem of uneven background image is basically solved and the uneven stitching gap completely disappeared. The image contrast improves 59.78% averagely. It greatly improves the accuracy and efficiency of the system to detect detection.Finally, according to the overall system-level experiments, we further validate that the image acquisition algorithms and algorithms for image background uniformity can improve overall system performance and efficiency after the realization on improved automatic leveling system, as well as the suitability for automated detection.